Beaverton, Oregon 97003

20276 SW Lela Lane

Beaverton, Oregon 97003

January 3, 2017

CRSO EIS

P.O. Box 2870

Portland, Oregon 97208 - 2870

Email:

Dear Sirs:

This letter is submitted in accordance with the federal agencies invitation to the public to submit comments on alternatives for long-term Columbia River System Operations (CRSO). The proposed alternative in this letter will simultaneously improve fish passage of downstream migrants, increase service life of existing turbine-generating equipment, and both improve power generation and increase revenue from power operations. The basic technology involved has been known for decades. However, recent technical developments and innovations have reduced its labor time in the field and total cost by about five-fold.

First, to document my expertise to address the issue, I have a Master of Science degree in fluid mechanics and hydraulics, am a registered Professional Engineer in both Oregon and Washington, have been a hydropower engineer for 50 years, have published 33 technical papers and a college textbook on hydropower engineering, have worked on every federal hydroelectric project in the Pacific Northwest, worked for the Corps of Engineers, BPA, and TVA, and am presently a professor at the Oregon Institute of Technology (OIT) in Wilsonville, teaching hydropower engineering and fluid mechanics.

Some alternatives to long-term system operations, seen recently in the literature, can be immediately eliminated from this solicitation. These concern dam breaching, or more precisely, removing the hydropower dams on the Snake River. That would significantly weaken and degrade the electrical grid in the Pacific Northwest, which is one of the finest, most reliable, and especially the most stable in the world. This stability is because of the large number of relatively slow rotating and extremely heavy hydroelectric generators. This massive amount of mechanical rotating or angular momentum can and does almost instantly absorb and supply large amounts of electrical power, thereby mitigating the effects of electrical transients.

Other generating resources, including renewables, have relatively high speed, light weight generators and do not contribute as much to grid stability. In other words, our grid can presently absorb a large amount of fluctuating additional electrical energy from renewable resources, but only if there is also sufficient hydroelectric generation on the line. If such hydro generation is eliminated from our grid we can utilize less and less alternative generating resources, or else degrade the stability and reliability of our grid. The proof of this is the manner in which the hydroelectric generators of some Corps of Engineer's projects are left connected to the grid to "motor" at synchronous speed with their turbines spinning in air. This is called "synchronous condensing," and is done to correct the power factor, i.e. voltage stability of the grid.

Therefore, we really only have a choice to either fix part of the present system's method of operation that we have ignored for decades or else keep on spilling to: minimize downstream fish mortality, forego potential additional revenue and waste one of our greatest natural resources.

One solution to our long-term system operational needs has existed since the first Kaplan turbine went into commercial service over a century ago. This involves the "index testing" of Kaplan turbines and its latest technical developments.

First, index testing, Type 1 efficiency optimization, is not an alternative. The term alternative implies something that substitutes for or replaces. Indexing is not proposed as a replacement for any aid to downstream fish migration, such as fish screens. It is an added alternative aid. Indexing is essential a "tune-up" of a Kaplan turbine to maximize its operating efficiency.

A comprehensive understanding of this technology involves knowledge of how hydraulic reaction turbines function in general, and specifically how those in the Columbia and Snake Rivers are operated. This indexing technology applies to both Kaplan and Francis type turbines, but the greater benefit is derived from its application to the CRPS Kaplan type turbines. All hydraulic reaction turbines physically operate by imparting a whirl component to the flow in their spiral or semi-spiral case and then the turbine runner (rotating element) straightens this flow out. This mechanical change of fluid angular momentum induces a torque in foot-pounds on the turbine shaft. The rate at which this change is done becomes foot-pounds per second, which is mechanical power.

The US Army Corps of Engineers has more than 100 Kaplan turbines in the FCRPS. These low head turbines, discharge the greatest flow per MW and are the turbines involved in downstream fish passage. These Kaplan's are vertical shafted, variable pitch, propeller type turbines developed by an Austrian Professor, Dr. Viktor Kaplan, in the 1920's. He modified a standard fixed blade propeller turbine so that the pitch angle of the blades could be varied continuously while the machine was in operation.

The head (difference in upstream and downstream elevations or pressure) and wicket gate opening (valve or vane opening) determine the individual turbine's volumetric flow rate and the unique angle at which the corresponding water velocity approaches the leading edges of the turbine's blades. The pitch angle of the blades needs to be continually adjusted to match that ever changing angle of this flow velocity vector in order to maintain maximum efficiency. At maximum efficiency, there is the least chance that a downstream migrant fish would experience blade strike, grinding or abrasion. Since these three are among the five highest methods of fish mortality in Kaplan turbines, operating at maximum efficiency will achieve the highest total survival of downstream migrating fish for any head and power output.

The basic problem with fish mortality on the Columbia and Snake River systems is that very few of the Corps' Kaplan turbine governors are able to properly control their turbines to continuously achieve this optimum blade angle. This is due to three factors:

a. every turbine, even those manufactured under the same contract (same manufacturer), are hydraulically different from every

other turbine and therefore the blade pitch angle for optimum efficiency varies from every other turbine for the same head and wicket gate opening,

b. less than one out of every ten of the Corp's Kaplan's on the Columbia River system is ever field tested to learn its true, unique and optimum blade angle, and

c. most Corps of Engineer governors have dead band limits on blade positioning errors that are many times larger than allowed by industry specifications for commercial governors.

The method to conduct a relative efficiency field test or index test to determine "a" above was developed in the 1920's by a New York utility. All Kaplan turbines in the Columbia River system are instrumented and have the embedded piping needed for index testing.

Historically, this testing does involve some field work and financial expense. The budgetary figure in 2006 for a Kaplan index test at a single head was $25k, which is why in "b," so few turbines are ever index tested at even one head during their roughly 70 year operating life. This is despite the fact that it has been repeatedly demonstrated that index testing pays for itself by the additional electrical energy produced in an extremely short time and extended operating life before refurbishment is needed. Further, several fish survivability tests conducted on newly indexed and efficiency optimized Kaplan turbines have found 48 hour hold time mortality rates do not exceed 2 - 3%. This mortality is lower than the mortality rates, under the same test conditions, of even the best powerhouse fish collection systems or the loss from being discharged through the gates of a spillway.

About 25 years ago, a device, named an Index Test Box (ITB), was invented that was capable of performing unattended index tests. That is, it was operated by simply plugging it into a governor and recording performance data while the generating unit remained in normal operation, utilizing a new Constant Power Testing Method that does not cause wide swings of flow and power, like conventional index testing methods. Bonneville Power Administration (BPA) evaluated this device, tested it at Portland General Electric's (PGE's) Hydro Plant #2 and found it worked perfectly. BPA then offered to purchase and supply one for every Corps of Engineers' hydropower plant in the FCRPS. The offer was rejected.

One of the proprietary features in the software of the ITB is a subroutine called, "Steady State." This subroutine waits until all individual data inputs are simultaneously steady-state (within programmable specified ramp rates) before recording a data point. Otherwise recorded data points simply form a data cloud. It is understood the Corps has attempted over the years to develop its own version of the ITB, but no test results have ever been published.

In Figures 1. and 2. below, data collected with the Corps of Engineers version of an ITB, called a Gate Blade Optimizer (GBO), was acquired by a Freedom of Information Act request and analyzed by comparing gate-blade positioning to the 3-D Cam surface programmed into the governor. The mechanism of the 3-D Cam actuator developed by HDC causes unpredictable position error as shown.

Figure 1.

Figure 2.

Meanwhile, separately over the intervening years and cumulating in last year's development, the inventor of the ITB perfected a method of achieving the most accurate of index testing results, without even having to actually conduct any field tests. This newest technique has been demonstrated at a privately owned, single unit Kaplan powerhouse, that uses the forebay at Dornea Dam in Oregon, owned by the Corps of Engineers. This method needs only the data normally recorded on the station's SCADA (Supervisory Control and Data Acquisition) system. A software program from the Index Test Box is used to sort and correlate this data into conventional tabular format and then a separate software program, developed by students at OIT (Oregon Institute of Technology in Wilsonville), analyses this data exactly as if it had been measured by field index tests.

On the following pages are a few graphs, from Corps of Engineers projects that demonstrate the foregoing principles.

Figure 3.

In Figure 3. above, the absolute efficiency profiles for the first 5 original turbines at The Dalles Dam are shown. These were determined by current meter tests in the late 1960's. They show that even though these units are nominally the same, that is made by the same manufacturer, under the same contract, and using the same model derived cam curve, they perform as completely different units. The difference in peak efficiency is about 9% and even the difference in the location of peak efficiencies occurs over several megawatts. In other words, the governor of each and every turbine needs to be programmed with its own unique, optimized individual cam curve derived from index tests if that powerhouse it to maximize its total efficiency and the survivability of downstream migrant fish.

Figure 4.

Figure 4. above shows the results of a traditional Kaplan turbine field index test utilizing 8 fixed blade angles. On the upper curves the location of peak efficiency, if the optimum individual cam curve were installed in this machine, is noted as 129 MW. The lower curves are just part of the data reduction technique. However, the real note on this graph is the other, lower brown "tangent line." This is actually the line showing how the machine is operating on the cam permanently installed in its governor, at the time of the test. At the time of the test, at 104 MW the relative efficiency dropped from an optimum of 547,000 to 536,000, a change of 2%.

That 2% did not simply disappear, but the lost energy it represents had to be destroyed within the fluid column before or at its discharge from the draft tube exit. If it were not destroyed, the tailwater would have to exist at two different elevations, an obvious impossibility. 104 MW is equal to 139,500 horsepower, and 2% of that is 2,790 horsepower. In other words, 2,790 horsepower (the power roughly equivalent to the combined power plants in two Abrams Mark 1, A1, main battle tanks) must be destroyed in heat, turbulence and vibration within a draft tube distance of about 100 feet. It is no wonder that downstream migrant fish emerging from a draft tube exit of a FCRPS Kaplan turbine, that is not operating at its best efficiency, will be completely disoriented and subject to predation.

In Figure 5. below from the same traditional field index test as Figure 4, which was conducted as accurately as the technique at that time allowed, produced the cam curve shown. First, the black line is the optimum cam curve determined by the index test points of tangency (the dark blue dots). It is noted there is a very slight scatter to the blue dots which is typical of the uncertainty of the traditional method of field index testing.

The real point of note is that the "as found" cam curve (in the light blue line) shows the permanently installed cam at the time of the test. Its wicket gate

servomotor has a lift off point at gate opening 10.5% smaller than the optimum, and the index test determined the permanently installed cam had a maximum blade positioning error of 3.5 degrees, too steep. Since this machine only has a total blade movement of 15 degrees, this is a blade positioning error of 23.3%! This result of a normal field index test on a Corps of Engineer Kaplan turbine is not unusual, being in fact typical of my experience in testing Corps Kaplan's.

(Continued on Next Page)

Figure 5.

The same format for an index test derived cam curve from a different test on a different Kaplan turbine at a different Corps Columbia River powerhouse is shown below in Figure 6. The permanently installed (or as found cam) is shown in dark blue dots with a dark blue line and the optimum cam curve derived from the index test is shown with pink dots and a pink line. The difference shows they both have about the same lift off point, but the "as found" cam positions the blades about 2 degrees flatter for the same wicket gate servomotor stroke or about a 4.5% wider wicket gate opening for the same blade angle. The green line, labeled verification test, is from a post index (or check) test to verify the governor is following the replacement cam within its tolerance band. By design, Corps of Engineers governors have a programmable deadband error on blade positioning to de-sensitize the governor closed-loop to prevent an unstable behavior. This causes excessive blade positioning error and increased turbulence and shear forces in the water column that are harmful to fish.